ADVANCED TECHNOLOGIES FOR IMAGE AND SIGNAL PROCESSING (ATISP)
Code: UR17ES41
Presentation :
This research unit named : Advanced Technologies for Image and Signal Processing ‘ATISP’, was founded in January 2017 at the National School of Electronics and Telecommunications of Sfax under the code UR17ES41.It works on advanced technologies in the field of signal and image analysis. The research program pursued within ATISP has made it possible to develop several scientific and technical skills, in particular in the following disciplines (axes):
Speech and audio signal processing;
Remote sensing and satellite imagery
The originality of the research program proposed in this unit consists in the creative association of interdisciplinary information allowing the global optimization of modern metrological tools (measurement systems) which can only be designed by an additive compilation of knowledge from various fields, such as medicine, digital signal processing, remote sensing, etc.
The general objective of the proposed research program is to develop approaches and advanced signal processing methods, using new technologies in electronics and computer science, to improve the quality of interpretation and analysis of data in medical imaging ‘for Computer Aided Diagnosis 'CAD'), as well as in speech and audio processing, in remote sensing and in embedded systems in biometrics etc...
This program concern multidimensional signal processing, image processing in very specific fields, remote sensing and the design of dedicated software and hardware platforms.
The projected ATISP research unit aims at various specific objectives related to advanced signal and image processing methods using cutting-edge technologies. The specific objectives for each application area can be summarized as follows :
Image processing : Computer-aided diagnosis (CAD) is a rapidly expanding field of research that offers very remarkable analysis and diagnostic and decision support possibilities :
Development of detection, segmentation and fusion algorithms to aid in computer-assisted diagnosis and therapeutic evaluation for the fight against cancer.
Development of detection, segmentation and fusion algorithms to detect certain malformations of the tumor type.
Development of fusion algorithms between various imaging modalities and audio. The primary goal is the complementarity for recognition purposes.
Speech and audio processing : This is an emerging field that is growing very rapidly in the industrial and medical domains.
Development of advanced recognition systems for phonemes, words and continuous speech.
Development of sophisticated speech analysis and synthesis systems.
Analysis of the voice signal with a view to estimating the formants by different techniques such as linear predictive coding (LPC), cepstrum, etc…
Automatic speech recognition in small and large vocabulary using statistical classification methods (Hidden Markov Models) and neural networks.
Statistical learning by specific algorithms in order to increase the discriminating capacities of HMMs.
Synthesis of the speech signal by various techniques, in particular statistical techniques.
Remote sensing and satellite imagery processing : Remote sensing is a high technology field which currently offers very important services in such domains as agriculture, geography and weather :
Development of algorithms for disaggregation and unmixing of series of optical and infrared satellite images in order to find the properties of the different classes of land cover at high resolutions.
Development of algorithms for merging images from several sensors with a view to complementarity, thus allowing better detection and monitoring of surfaces.
Development of approaches for inverting LiDAR tree data to estimate structural and biophysical properties.
